BMC Microbiology BioMed Central

Research article Open Access Carbon and arsenic metabolism in Thiomonas strains: differences revealed diverse adaptation processes Christopher G Bryan†1,4, Marie Marchal†1, Fabienne Battaglia-Brunet2, Valérie Kugler1, Christelle Lemaitre-Guillier3, Didier Lièvremont1, Philippe N Bertin1 and Florence Arsène-Ploetze*1

Address: 1Génétique Moléculaire, Génomique et Microbiologie, UMR 7156 CNRS and Université de Strasbourg, 28, rue Goethe, 67000 Strasbourg, France, 2BRGM, Environnement et Procédés, Unité Ecotechnologie, Avenue Claude Guillemin, 45060 Orléans, France, 3Plateforme Protéomique, IFR 1589 CNRS, 15 rue René Descartes, 67084 Strasbourg, France and 4Current address: Centre for Bioprocess Engineering Research, Department of Chemical Engineering, University of Cape Town, Rondebosch 7701, South Africa Email: Christopher G Bryan - [email protected]; Marie Marchal - [email protected]; Fabienne Battaglia- Brunet - [email protected]; Valérie Kugler - [email protected]; Christelle Lemaitre-Guillier - [email protected]; Didier Lièvremont - [email protected]; Philippe N Bertin - [email protected]; Florence Arsène- Ploetze* - [email protected] * Corresponding author †Equal contributors

Published: 23 June 2009 Received: 25 March 2009 Accepted: 23 June 2009 BMC Microbiology 2009, 9:127 doi:10.1186/1471-2180-9-127 This article is available from: http://www.biomedcentral.com/1471-2180/9/127 © 2009 Bryan et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract Background: Thiomonas strains are ubiquitous in arsenic-contaminated environments. Differences between Thiomonas strains in the way they have adapted and respond to arsenic have never been studied in detail. For this purpose, five Thiomonas strains, that are interesting in terms of arsenic metabolism were selected: T. arsenivorans, Thiomonas spp. WJ68 and 3As are able to oxidise As(III), while Thiomonas sp. Ynys1 and T. perometabolis are not. Moreover, T. arsenivorans and 3As present interesting physiological traits, in particular that these strains are able to use As(III) as an electron donor. Results: The metabolism of carbon and arsenic was compared in the five Thiomonas strains belonging to two distinct phylogenetic groups. Greater physiological differences were found between these strains than might have been suggested by 16S rRNA/rpoA gene phylogeny, especially regarding arsenic metabolism. Physiologically, T. perometabolis and Ynys1 were unable to oxidise As(III) and were less arsenic-resistant than the other strains. Genetically, they appeared to lack the aox arsenic-oxidising genes and carried only a single ars arsenic resistance operon. Thiomonas arsenivorans belonged to a distinct phylogenetic group and increased its autotrophic metabolism when arsenic concentration increased. Differential proteomic analysis revealed that in T. arsenivorans, the rbc/cbb genes involved in the assimilation of inorganic carbon were induced in the presence of arsenic, whereas these genes were repressed in Thiomonas sp. 3As. Conclusion: Taken together, these results show that these closely related bacteria differ substantially in their response to arsenic, amongst other factors, and suggest different relationships between carbon assimilation and arsenic metabolism.

Page 1 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

Background prised Thiomonas cuprina, T. intermedia, T. perometabolis Microorganisms play an essential role in shaping the nat- and T. Thermosulfata [17,18]. Thiomonas perometabolis was ural environment. They have evolved specific metabolic isolated from soil at a building site in Los Angeles, U.S.A., pathways allowing them to utilise a wide range of sub- as Thiobacillus perometabolis [19]. It was differentiated from strates, many of which are toxic to higher organisms. Thiobacillus intermedius (now T. intermedia, the type spe- Through the conversion of both anthropogenic and natu- cies of the genus) as it was apparently unable to grow rally occurring pollutants to less toxic products, such autotrophically. However, Katayama-Fujimura and Kurai- microorganisms effect widespread natural bioremedia- shi [20] have since suggested that this is not true. Recently tion. An important toxic compound is arsenic, a metalloid described species include Thiomonas. arsenivorans [21] and that can cause multiple health effects including diabetes, the Thiomonas strains 3As [12], Ynys1 [22] and WJ68 [14]. hypertension, skin lesions and skin and internal cancers Thiomonas sp. 3As was obtained from the Carnoulès mine [1]. Arsenic occurs in soils and water bodies both natu- tailings, Southern France [12]. It was shown that this bac- rally and as a result of anthropogenic processes. A major terium could gain energy from the oxidation of arsenic. anthropogenic source is the mining industry, where the The presence of carboxysomes and the detection of the processing of sulfide ores produces large quantities of cbbSL genes encoding ribulose 1,5-bisphosphate carboxy- sulfidic wastes which may be rich in arsenic-bearing com- lase/oxygenase, led the authors propose that this strain pounds such as arsenopyrite. The weathering of these may be able to fix CO2. T. arsenivorans was isolated from minerals leads to the formation of acid mine drainage another arsenic-rich mine residue at the Cheni former (AMD), generally characterised by elevated sulfate, iron gold mine, Limousin, France [21]. The Cheni site is not and other metal concentrations [2], and thus the transport very acidic (pH ~6.0), but is highly contaminated with of many toxic elements such as inorganic forms of arsenic, arsenic (6.0 mg g-1 in the solid phase and ~1.33 mM in the arsenite (As(III)) and arsenate (As(V)). This often results liquid phase) [23]. T. arsenivorans has been shown to oxi- in chronic and severe pollution of the surrounding envi- dise arsenic and ferrous iron, and is able to grow ronment, with a substantial reduction of the indigenous autotrophically using arsenic as the sole energy source biota. [21]. Strain Ynys1 was isolated from ferruginous waters which have been draining from an adit since the closure Numerous arsenic-oxidising microorganisms, especially of several coal mines near to the village of Ynysarwed, Proteobacteria, are able to oxidise As(III) to As(V) in order Wales, U.K. [22]. The waters were of relatively neutral pH to detoxify their immediate environment. This biological (pH 6.3) with elevated iron loading (300 mg L-1) and have As(III) oxidation is of particular importance, As(III) being led to significant pollution of the area [22]. Strain WJ68 more soluble and more toxic than As(V) [3]. Additionally, was the dominant isolate obtained from effluent draining in acidic environments such as those impacted by AMD, all three of the compost bioreactors of a pilot-scale biore- natural remediation can occur as a result of the concurrent mediation plant receiving water from the Wheal Jane tin oxidation of ferrous iron and arsenite, leading to the mine, Cornwall, U.K. [14]. Both WJ68 and Ynys1 are coprecipitation of both [4]. Therefore, understanding fac- known to oxidise ferrous iron, while WJ68 has been tors that influence the competitiveness, diversity and role shown to oxidise arsenite [15]. of these organisms is an essential step in the development of bioremediation systems treating arsenic contaminated These five strains are interesting in terms of arsenic metab- environments. olism: T. arsenivorans, WJ68 and 3As are able to oxidise As(III), while Ynys1 and T. perometabolis are not. Moreo- Certain bacterial strains are able to use arsenite as an elec- ver, T. arsenivorans and 3As present interesting physiolog- tron donor. By gaining energy, as well as removing the ical traits, in particular that these strains are able to use more toxic arsenic species, such bacteria may gain an As(III) as an electron donor. However, differences advantage over other microorganisms [5]. Arsenite oxi- between Thiomonas strains in the way they have adapted dase, the enzyme catalysing As(III)-oxidation, has been and respond to arsenic have never been studied further. well characterised in several bacterial strains [6-11]. An The connection between carbon and arsenic metabolism important group of As(III)-oxidising bacteria belong to in these strains, particularly inorganic carbon assimilation the Thiomonas genus, and are ubiquitous in arsenic-con- and arsenite as energy source, has never been compared. taminated environments [12-15]. Thiomonas strains are Therefore, analysis was undertaken to examine these able to gain energy from the oxidation of reduced inor- physiological aspects in these five Thiomonas strains. ganic sulphur compounds (RISCs) [16], and are defined as facultative chemolithoautotrophs which grow opti- mally in mixotrophic media containing RISCs and organic supplements. These bacteria are also capable of organotrophic growth [17]. The original description com-

Page 2 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

Results perometabolis. The aoxAB genes of WJ68 were much more Phylogenetic, phenotypic and genotypic analyses of the divergent than those of T. arsenivorans and 3As (data not five Thiomonas strains shown). This is in agreement with previous findings Phylogenetic analyses of amplified 16S rRNA and rpoA showing that the aoxB gene of WJ68 groups neither with gene products confirmed the occurrence of two distinct T. arsenivorans nor the Group I thiomonads [10], monophyletic groups as had been suggested previously (Quéméneur, personal communication). The inability of [15]. SuperGene analysis (Figure. 1A) was performed T. perometabolis and Ynys1 to oxidise arsenite further using concatenated 16S rRNA and rpoA gene sequences of implied that the aox operon was absent in these strains. each strain. These results placed T. perometabolis with WJ68 and Ynys1. Along with Thiomonas sp. 3As, these The MIC of As(III) for strains 3As, WJ68 and T. arsenivo- strains grouped together in Group I, while T. arsenivorans rans was 10 mM, higher than for strains Ynys1 and T. per- was part of Group II. ometabolis (Table 1). Additionally, strain Ynys1 was more sensitive to As(V) than the other strains. Arsenic resistance Various tests were carried out to examine the physiologi- in bacteria is in part due to the expression of aox genes but cal response of the five strains to arsenic. This was coupled also of the ars arsenic-resistance genes [8]. Among these, with a PCR-based approach to determine the presence of arsC encodes an arsenate reductase and arsA and arsB genes involved in arsenic metabolism. In agreement with encode an arsenite efflux pump. Analysis of the Thiomonas previous data, strains 3As, WJ68 and T. arsenivorans oxi- sp. 3As genome (Arsène-Ploetze & Bertin, unpublished) dised arsenite to arsenate in liquid media whereas T. per- revealed the presence of two copies of the arsB gene, ometabolis and Ynys1 did not (Table 1). The aoxAB genes denoted arsB1 and arsB2. These genes were found to be encoding the arsenite oxidase large and small subunits of distantly related, sharing just 70.2% sequence identity. In Thiomonas sp. 3As and T. arsenivorans have previously order to compare the occurrence, copy number and type been characterised [12,24]. Positive PCR results using of ars genes present in the different Thiomonas strains, PCR primers which targeted a region of the aoxAB genes were amplifications using generic arsB primers were performed. obtained with DNA from all strains except Ynys1 and T. As expected, RFLP and sequence analysis confirmed the

arsB1 A B

Ralstonia eutropha H16 At. caldus Ynys1

3As Tm. arsenivorans L. ferriphilum Tm. perometabolis SuperGene 3As (16S rRNA:rpoA hyy)brid) At. caldus transposon 100 Tm. perometabolis

Tm. arsenivorans 52 0.03 Ynys1 3As 010.1 100 Hm. arsenicoxydans 2 Hm. arsenicoxydans 1 WJ68 arsB2

PhylogeneticinFigure this study 1 dendrogram of the SuperGene construct of both the 16S rRNA and rpoA genes (A) of the Thiomonas strains used Phylogenetic dendrogram of the SuperGene construct of both the 16S rRNA and rpoA genes (A) of the Thi- omonas strains used in this study. Ralstonia eutropha H16 served as the outgroup. Numbers at the branches indicate per- centage bootstrap support from 500 re-samplings for ML analysis. NJ analyses (not shown) produced the same branch positions in each case. The scale bar represents changes per nucleotide. (B) Phylogenetic dendrogram of the arsB genes of the Thiomonas strains used in this study and some other closely-related bacteria. Both ML and NJ (not shown) analysis gave the same tree structure. The scale bar represents changes per nucleotide. Sequences obtained using the arsB1- and arsB2-specific internal primers were not included in the analysis as the sequences produced were of only between 200 – 350 nt in length.

Page 3 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127 Table 1: Summary of physiological physiological of Table 1:Summary presence of 1.33 mM of arsenite: "+" indica colony forming units (cfu) observed (cfu) units forming colony a extract (YE), thiosulfateextract orarse except for WJ68,tested in Tm. perometabolis Tm. Diameter (mm) of swi Tm. arsenivorans Thiomonas Thiomonas strains ns 2556-(5)-+--n + dnd nd - +++ - nd - - + - (35%) - 5.6 12.5 5 - Ynys1 J8+1 0 87+(% d+ + dnd nd - +++ ++ nd - - + + + (6%) 38.7 > 100 10 + WJ68 A 01029/+++-++ + 6% - (67%) - (69%) - +++ ++ + - + + + / 2.9 100 10 + 3As Oxidation mming ringmming formed0.3% onagar plates afterh 72 incu As(III) As(III) 01045+(4)+-++ ++ + ++(5)/ +(25%) ++ +++ ++ ++ ++ + - + (24%) + 4.5 100 10 + 0 d-++-n nd nd - +++ - nd - - + - / 0 >100 5 - 0.5% YE, without As(III). As(III). without YE, 0.5% nite or combinat or nite sII As(V) As(III) after10days;-,noincrease; I m)Motility (mM) MIC and genetic data data genetic and tes a diameter of swimmingtes adiameter ri rei-eae hntp/eoyeGrow Arsenic-related phenotype/genotype ions thereof. g 1.33 mM As(III) in MCSM. nd: nodata. nd: inMCSM. As(III) 1.33 mM obtained forthe a d 5,33 mMincaseof 3As, WJ68, and f Tested with 0.1, 0.2, 0.3% or 0.5% YE in on strain motility Effect of Effect arsenite ng greaterng than inabsencearsenite, of "-" asmaller o b Thiomonas bation expressed as adifference withmotile non strai oA rB arsB2 arsB1 aoxAB arsenic-related genes PCR amplification of PCR amplification strains usedinthisstudy. Tm. arsenivorans As(III) absence of As(III), with 0.1, 0.2 or 0.3% YEabsence0.3%0.2 ofAs(III), or with0.1, d, e , 2.67 mM in ca in , 2.67mM YE +As(III) YE th withdifferent electrondonors ne and "/" no change. f se of Ynys1 andse of ns (forming colonies of < 3 mm diameter); mmdiameter); < 3 colonies of (forming ns YE f YE +S 2 O Tm. perometabolis c Basel medium (MCSM or m126) or medium(MCSM Basel 3 and 1.3 mM of As(III), or with 0.3% YE an YE 0.3% with or As(III), of mM 1.3 and 2- c S 2 O 3 2-e . e Growth is expressed as an increase of asanincrease Growth isexpressed YE 0.1YE gL Influence of As(III) on final cell final on As(III) of Influence b Motility was tested in the Motility wastestedin concentration -1 amendedwith either yeast d 2.6 mM As(III), mMAs(III), d 2.6 YE 0.2 g L 0.2 YE g -1

Page 4 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

presence of the arsB1 and arsB2 genes in strain 3As (Table All the physiological and genetic analyses revealed that 1). In contrast, only the arsB1 gene could be detected the response to arsenic differed in the five Thiomonas using DNA from T. perometabolis, Ynys1 and WJ68, even strains; some of these differences were correlated with dif- when internal primers specific for the arsB2 gene were ferences in the genetic content. used. Conversely, only the arsB2 gene was detected in T. arsenivorans. As(III) as an energy source, and the fixation of carbon dioxide The phylogeny of the arsB1 and arsB2 genes was analysed, Only T. arsenivorans, 3As and WJ68 were able to grow in excluding the sequences obtained using the arsB2 internal basal media with yeast extract as the sole energy source primers that were too short. The arsB2 gene sequence for (Table 1). During these growth experiments, soluble sul- strain 3As was taken directly from the annotated genome fate concentrations remained the same or decreased (Arsène-Ploetze & Bertin, unpublished). The data showed slightly (data not shown), indicating that energy was that while they are all related to the arsB genes of Lept- gained from the oxidation of compounds other than any ospirillum spp. and Acidithiobacillus caldus, the type 1 and trace RISCs in the yeast extract, most probably organic car- type 2 genes formed two very distinct clades and have bon. These observations suggest that all strains except clearly diverged at an evolutionarily distant point in time Ynys1 and T. perometabolis are organotrophic. All strains (Figure. 1B). were able to grow in the presence of YE and thiosulfate (Table 1). In these thiosulfate-amended cultures, sulfate The motility of Herminiimonas arsenicoxydans, an arsenic- concentrations increased following incubation (data not oxidising bacterium is greater in the presence of arsenite shown), indicating that thiosulfate had been oxidised. [25]. Motility tests revealed that the five Thiomonas strains This suggests that all strains were able to use this RISC as reacted differently to the metalloid (Table 1). Strain T. per- an energy source and are therefore chemolithotrophic. In ometabolis was found to be non-motile irrespective of all cases, greater growth occurred in thiosulfate-amended arsenite concentrations. Among the motile strains, three cultures, suggesting that mixotrophic conditions are opti- distinct phenotypes were observed: those for whom mal for the growth of these strains. It was however motility was not affected by arsenite concentration (strain observed that T. arsenivorans grew better in MCSM liquid 3As); those who showed increased motility with increas- medium, whereas T. perometabolis and Ynys1 grew better ing arsenite concentrations (strains T. arsenivorans and in m126 medium (3As and WJ68 grew equally well in WJ68) and those who showed decreased motility with both; data not shown). MCSM contains 2 times less thio- increasing arsenite concentration (Ynys1). WJ68 was three sulfate and suggests that the optimal thiosulfate concen- to four times more motile than all of the other strains. A tration is lower in the case of T. arsenivorans. concentration of 2.67 mM arsenite appeared to have an inhibitory effect on T. arsenivorans and WJ68 motility Only T. arsenivorans was able to grow in basal media with- (data not shown). out yeast extract with either thiosulfate or arsenite as the sole energy source (Table 1). Although direct cell enumer- ation of T. perometabolis cultures was not possible due to its propensity to form flocs during growth, no growth, 5.0 flocular or otherwise, was observed in the YE-free media. 4.5 The growth of T. arsenivorans was stimulated by 1.33 mM 4.0 As(III) in presence of 0.1 g L-1 yeast extract, but this posi- )

-1 3.5 L tive effect was no longer detected in presence of 0.2 g L-1 3.0 (mg yeast extract. The ability of T. arsenivorans to grow 2.5

fixed autotrophically using As(III) as the sole energy source was n o o 2.0 confirmed by the observation of increasing quantities of 1.5 Carb carbon fixed as more As(III) was oxidised (Figure. 2). This 1.0 demonstrated that T. arsenivorans was able to use energy 050.5 gained from the oxidation of As(III) to fix inorganic car- 0.0 bon. In contrast, strain 3As was unable to fix inorganic 0 200 400 600 800 1000 Arsenite oxidised (mg L-1) carbon under the same conditions (in MCSM), as 1.33 mM As(III) was found to inhibit growth in presence of 0.1 CarbonransFigure 2fixed as a product of As(III) oxidised by T. arsenivo- or 0.2 g L-1 yeast extract (Table 1), and this strain was una- Carbon fixed as a product of As(III) oxidised by T. ble to grow in presence of As(III) as the sole energy source. arsenivorans. Error bars, where visible, show standard devi- ation; n = 3 for each data point. Figure 2 shows an essentially linear relationship between carbon fixed and arsenic oxidised, corresponding to 3.9

Page 5 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

Table 2: Arsenic-induced or repressed proteins in T. arsenivorans and Thiomonas sp. 3As.

Functional class Metabolic pathway Gene Protein Induction/repression by Asa

T. arsenivorans Thiomonas sp. 3As

Energy and carbon Calvin Cycle rbcL Ribulose-1,5-bisphosphate carboxylase/ +- metabolism oxygenase large subunit cbbFC1 Fructose-1,6-bisphosphatase + 0 cbbA1 Fructose biphosphate aldolase 0 -

TCA cycle/reductive icd Isocitrate dehydrogenase, specific for +0 carboxylate cycle NADP+

Glyoxylate and aceB Malate synthase A + 0 dicarboxylate metabolism gltA Citrate synthase + 0 aceA Isocitrate lyase 0 + / Tartrate dehydrogenase/decarboxylase 0+ (TDH) (D-malate dehydrogenase [decarboxylating])

Glycolyse/gluconeogenesis ppsA Phosphoenolpyruvate synthase + - aceE Pyruvate dehydrogenase E1 component + - lpdA Dihydrolipoyl dehydrogenase +0 (Pyruvate dehydrogenase E3 component) eno2 Enolase 0 -

Thiosulfate oxydation / Putative sulfur oxidation protein SoxB 0 -

Cellular processes, Arsenic resistance arsA2 Arsenical pump-driving ATPase + 0 transport and binding proteins

arsC1 Arsenate reductase 0 +

High temperature hldD ADP-L-glycero-D-manno-heptose-6- +0 resistance epimerase

General stress groL GroEL, 60 kDa chaperonin + 0

Other stresses ahpF Alkyl hydroperoxide reductase subunit F 0 -

Twitching/motility/ / Putative methyl-accepting chemotaxis 0- secretion protein

/ Putative type IV pilus assembly protein PilM 0 -

Cell division / Putative cell division protein 0 -

DNA metabolism, DNA bending, supercoiling, gyrA DNA gyrase subunit A + - transcription and inversion protein synthesis

RNA degradation pnp Polyribonucleotide nucleotidyltransferase + -

Protein synthesis fusA Elongation factor G (EF-G) + 0 tufA Elongation factor Tu + 0 rpsB 30S ribosomal protein S2 + 0 rpsA 30S ribosomal protein S1 0 -

a + and -: these proteins are more or less abundant in the presence of As(III), respectively. 0: no difference observed (for details, see Additional File 1).

Page 6 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

mg C fixed for 1 g of As(III) oxidised, i.e. 0.293 mg C fixed each strain (Figure. 3, Table 2 and see Additional file 1). mM-1 As(III). It requires 40 J to produce 1 mg of organic In both strains, arsenic-specific enzymes (ArsA2 in T. carbon cellular material from CO2 [26]. The energy pro- arsenivorans, ArsC1 in 3As) were more abundant in the duced from the oxidation of As(III) with O2 is 189 J presence of As(III), suggesting that a typical arsenic-spe- mMol-1 [27]. As a consequence, if 100% of this energy was cific response occurred in both strains. ArsA2 is part of the used for carbon fixation, 4.73 mg C would be fixed for 1 efflux pump with ArsB2 and is encoded by the ars2 mM As(III) oxidised. Thus, in this experiment, 6% of the operon. Moreover, expression of a putative oxidoreduct- energy available from arsenic oxidation was used for car- ase (THI3148-like protein) was induced in the presence of bon fixation. This result is in accordance with the 5 to arsenic. This protein is conserved in At. caldus, with 90% 10% range of efficiency for carbon fixation by various amino-acid identity (Arsène-Ploetze & Bertin, unpub- autotrophic bacteria [26]. lished). The At. caldus gene encoding this THI3148-like protein is embedded within an ars operon. This protein is Enzymes involved in carbon metabolism and energy also conserved in more than 56 other bacteria, for exam- acquisition are expressed differently in T. arsenivorans ple in Mycobacterium abscessus (51% identity) and Lactoba- and 3As in response to arsenic cillus plantarum (48% identity). In these two cases the Protein profiles expressed in MCSM or m126 media, in corresponding gene was also found in the vicinity of ars the presence and absence of arsenic were compared in genes.

  ] Y ] Y ] Y ] Y            A                                          

`   

 

   B         

     

          `



      1 .Q% ^_ 8 IQ`^_

FigureDifferential 3 proteomic analysis in T. arsenivorans and Thiomonas sp. 3As strains, in response to As(III) Differential proteomic analysis in T. arsenivorans and Thiomonas sp. 3As strains, in response to As(III). On the gel presented are extracts obtained from (A) T. arsenivorans or (B)Thiomonas sp. 3As cultivated in the absence (left) or in the pres- ence (right) of 2.7 mM As(III). Spots that are circled showed significant differences of accumulation pattern when the two growth conditions were compared. Protein sizes were evaluated by comparison with protein size standards (BenchMark™ Protein Ladder, Invitrogen).

Page 7 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

The expression of several proteins involved in other met- as sensitive to As(III) as the other strains. Moreover, the abolic pathways changed, suggesting that in the presence inhibitory effect of arsenite on Ynys1 motility suggests a of arsenic, the general metabolism of T. arsenivorans and greater susceptibility of this strain to the metalloid. This 3As was modified. Indeed, enzymes involved in glyoxy- could be due to the absence of aox or ars genes. Indeed, late metabolism were more abundant in the presence of these two strains are unable to oxidize As(III), probably as arsenic, suggesting that expression of such proteins is reg- they lack aox genes. Moreover, arsB2 genes were not ulated in response to arsenic in both strains. However, detected in Ynys1 and T. perometabolis. Therefore, it is several changes observed were clearly different between probable that these two strains have only a single set of both strains. In T. arsenivorans, two proteins involved in arsenic resistance genes that can be expressed. Interest- CO2 fixation (ribulose-1,5-biphosphate carboxylase ingly, WJ68 was found to be equally resistant to arsenic as (RuBisCo) and fructose-1,6-biphosphatase) were more these strains, yet no arsB2 gene could be amplified by abundant when cells were grown in the presence of PCR. The same is true for T. arsenivorans, for which no arsenic, whereas in Thiomonas sp. 3As, such proteins were arsB1 gene was detected by PCR, yet it was again as resist- less abundant in the presence of As(III). In addition to ant as those strains shown to possess both the ars1 and these proteins, it was observed that enzymes involved in ars2 operons. One possible explanation is that WJ68 pos- major carbon metabolism (glycolysis, neoglucogenesis) sesses two copies of the ars1 operon and T. arsenivorans or energy metabolism (thiosulfate oxidation, oxidative has two copies of the ars2 operon. Alternatively, the phosphorylation) were less abundant in 3As in the pres- higher resistance capacities of T. arsenivorans, Thiomonas ence of As(III). This observation correlated with the phe- sp. 3As, and WJ68, as compared to Ynys1 and T. perome- notypic observation that the strain 3As grew better in the tabolis may be due to greater As(III) oxidation capacity of absence of arsenic (Table 1). these strains.

Discussion The arsenic response observed in T. arsenivorans and 3As Two groups could be distinguished within the Thiomonas revealed that the proteins involved in arsenic resistance strains studied: Group I comprises all the strains in this (ars genes) were more highly expressed in the presence of study except T. arsenivorans, which is part of a second arsenic, as shown previously for H. arsenicoxydans [25,28], group, Group II. As described by Moreira and Amils [17], Pseudomonas aeruginosa [29] and Comamonas sp. [30]. all of the strains grew better in mixotrophic media con- Therefore, such a feature seems to be a common arsenic taining both thiosulfate and organic supplements, and response. In H. arsenicoxydans, other proteins that were used RISCs as an energy source. This suggests that lithotro- shown to be more abundant in the presence of arsenic phy is a general characteristic of the Thiomonas genus. In were involved in oxidative stress, DNA repair and motil- contrast, neither strain Ynys1 nor T. perometabolis could ity. In this study, such proteins (hydroperoxide reductase, grow organotrophically in the absence of a reduced sulfur methyl-accepting chemotaxis protein, PilM) were induced compound, suggesting that, despite previous findings, fac- in Thiomonas sp. 3As whereas in T. arsenivorans, only gen- ultative organotrophy is not a general property of the Thi- eral stress proteins were induced. These observations sug- omonas genus. To improve our understanding of these gest that the response to the stress induced by arsenic important arsenic-resistant bacteria, several metabolic involves different regulatory mechanisms in 3As and T. and genetic properties were investigated. It appears that arsenivorans. Contrary to this arsenic-specific response, the much greater physiological differentiation regarding other arsenic-regulated proteins identified in the Thi- arsenic response was possible between these Thiomonas omonas strains did not share a similar expression pattern strains than may have been previously suggested. Clearly with other arsenic-resistant bacteria. Thus it appears that organisms that are phylogenetically close can differ greatly while there may be a common arsenic response between physiologically, in particular concerning specific meta- all the bacteria, the general metabolism may be differen- bolic traits such as the metabolism of arsenic. For exam- tially adapted to each environment from which these ple, the effects of arsenic on the motility of all strains strains originated. In particular, T. arsenivorans has unique appeared to be somewhat random, and cannot easily be traits in terms of arsenic, carbon and energy metabolism related to any of the phylogenetic or physiological data that distinguish it from the other strains examined. obtained. It is worth noting that both T. arsenivorans and WJ68 strains exhibited increased motility in the presence Thiomonas arsenivorans can grow autotrophically using of arsenic. This may indicate a potential energetic role of either As(III) or thiosulfate as the sole energy source. Sur- the element for these strains, as proposed for the arsenic- prisingly, the differential protein expression analysis oxidising bacterium, H. arsenicoxydans [25]. revealed that even in the presence of yeast extract, proteins involved in CO2 fixation through the Calvin-Benson-Bas- Other physiological divergences concern arsenic resist- sham cycle and enzymes involved in the glycolysis/ ance. Ynys1 and T. perometabolis were approximately twice neoglucogenesis were expressed. In addition, it was

Page 8 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

shown in the present study that T. arsenivorans induces Concentrations of 10, 5.0, 2.25, 1.25 and 0 mM As(III) or expression of carbon fixation-specific enzymes in the pres- 100, 50, 25, 12.5, 6.3 and 0 mM As(V) were tested at 30°C ence of arsenic. This observation was correlated with an for up to 10 days. The ability of each strain to oxidise increased CO2 fixation efficiency when arsenic concentra- arsenite was tested in triplicate, in liquid media amended tion increased. This suggests that an increase in cbb genes with 0.67 mM arsenite. Detection of As(III) and As(V) was expression in the presence of arsenic improves its capacity performed by inductively coupled plasma-atomic emis- to fix CO2. On the other hand, the opposite observation sion spectrometry (ICP-AES) as described by Weeger et al. was seen with Thiomonas sp. 3As. Therefore, the proteomic [33]. To test the ability of each strain to grow in the results obtained from the present study suggest that these absence of a reduced inorganic sulfur source or organic two Thiomonas strains react differently to their arsenic- carbon source, pre-cultures grown in standard media were contaminated environments. The other differences harvested by centrifugation at 10 K g for 10 min, washed observed concern DNA metabolism, transcription and and resuspended in a basal medium (m126 medium with protein synthesis. It appears that, in the presence of no thiosulfate or yeast extract). These were then used to arsenic, T. arsenivorans is still able to express proteins inoculate the test liquid media and incubated at 30°C for required for optimal growth whereas 3As is not. 10 days. Soluble sulfate concentrations were determined turbidimetrically by the formation of insoluble barium Conclusion sulfate, as described by Kolmert et al. [34]. Bacterial These observations revealed that carbon assimilation, growth in media containing YE was assessed using optical energy acquisition and arsenic metabolism of these strains density at 600 nm. Viable cell counts were used to meas- are linked. However, they do not share a common mech- ure growth in the media lacking YE, as described by Miles anism, since metabolisms required for growth and carbon and Misra [35] using appropriate gelled media, as the assimilation are stimulated in T. arsenivorans in the pres- autotrophic growth yield would be much lower. Where YE ence of arsenic, but repressed in Thiomonas sp. 3As. Fur- was omitted, the media contained either the normal con- ther work is needed to test if a common mechanism centration of thiosulfate or 5.33 mM arsenite (or 2.67 mM occurs to regulate carbon assimilation and arsenic for those strains sensitive to arsenite) as an electron response in other Thiomonas strains. However, to our donor. In the case of arsenite-amended media, pre-cul- knowledge, this is the first example of such a link between tures were grown in the presence of 2.67 mM arsenite. arsenic metabolism and carbon assimilation. To determine autotrophic growth yield as a product of Methods As(III) oxidised, triplicate cultures were grown in liquid Culture media MCSM without YE or thiosulfate containing either 0.66 or All strains except T. arsenivorans were routinely cultured 1.33 mM As(III), at 25°C in static conditions. To test con- on m126 (modified 126 medium) gelled or liquid centrations greater than 1.33 mM, initial cultures contain- medium. Medium m126 contains: (g L-1) yeast extract ing 1.33 mM As(III) were inoculated. As soon as the (YE; 0.5); Na2S2O3 (5.0); KH2PO4 (1.5); Na2HPO4 (4.5); As(III) had been oxidised, more As(III) was added from a MgSO4·7H2O (0.1); (NH4)Cl (0.3), adjusted to pH 5.0 concentrated (0.13 M) stock solution to a final concentra- with H2SO4 prior to sterilisation. T. arsenivorans was rou- tion of 1.33 mM. Once this had been oxidised, the process tinely cultured on a modified MCSM medium (MCSM) was repeated until the desired total quantity of As(III) had [31] with vitamins and trace elements omitted, yeast been added. The oxidation of As(III) to As(V) was ana- extract added to a final concentration of 0.5 g L-1 and lysed as described by Battaglia-Brunet et al. [31]. The pH -1 Na2S2O3 to a final concentration of 2.5 g L . Variations of was adjusted to pH 6.0 using a sterile NaOH solution these media included omitting yeast extract and/or thio- before each As(III) addition. Once all of the As(III) had sulfate. Where no yeast extract was included, trace ele- been oxidised, each culture was centrifuged at 10 kg for 15 ments were added, as described previously [32]. Where min and the pellet resuspended in 10 mL MCSM. The required, the media were gelled by the addition of 12 g L- total organic carbon concentration of this suspension was 1 agar (final concentration). Arsenite (As(III)) and arse- analysed using an OI ANALYTICAL 1010 apparatus nate (As(V)) were added to media to the desired concen- according to the AFNOR NF EN 1484 method. The influ- tration from sterile stocks of 667.4 mM of the metalloid ence of As(III) on final cell concentration in the presence ion in ddH2O, from NaAsO2 (Prolabo) and of an organic substrate was determined with strains 3As Na2HAsO4·7H20 (Prolabo) salts, respectively. and T. arsenivorans in MCSM complemented with 0.1 or 0.2 g L-1 yeast extract. Final cell concentration was deter- Physiological tests mined by measuring optical density at 620 nm. Minimum inhibitory concentration (MIC) experiments were performed using gelled media, amended with a Strain motility was assessed using growth media supple- range of concentrations of either arsenite or arsenate. mented with 0.3% agar as described previously [36].

Page 9 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

Three separate cell cultures of each strain were analysed in using a CapLC capillary LC system (Waters, Altrincham, triplicate. UK) coupled to a hybrid quadrupole orthogonal accelera- tion time-of-flight tandem mass spectrometer (Q-TOF Differential protein expression analysis Micro, Waters). Diluted sample (5 μL) was first loaded, T. arsenivorans and Thiomonas sp. 3As strains were grown concentrated and cleaned up onto a C18 PepMap precol- in MCSM and m126, respectively, with or without 2.7 umn cartridge (LC Packings) and then separated on-line mM As(III). Cells were harvested by centrifugation (7 K g, by the analytical reversed-phase capillary column 10 min, 4°C). Cell lysis was performed as described pre- (NanoEase C18, 75 μm i.d., 15 cm length; Waters) with a viously [37]. Proteins were precipitated using the 2-D 200 μL min-1 flow rate. The gradient profile used consisted Clean-up kit (Amersham Biosciences) and resuspended in of a linear gradient from 97% A (97.9% H2O, 2% ACN rehydratation buffer (364 g L-1 thiourea, 1000 g L-1 urea, and 0.1% (v/v) HCOOH) to 95% B (98% ACN, 1.9% -1 - 25 g L CHAPS, 0.6% (v/v) IPG buffer Pharmalyte, 10 g L H2O and 0.1% (v/v) HCOOH) for 45 min followed by a 1 DTT and 0.01% (w/v) bromophenol blue). Protein con- linear gradient to 95% B for 3 min. Internal calibration centration was determined using the 2-D Quant kit was assumed by the Lockspray module (Waters) that (Amersham Biosciences). switches to a reference source (leucine enkephalin M2+ = 556.2551 m/z) every 10 seconds during the acquisition Three hundred μg of this extract were loaded onto an 18 run. The spray system (liquid junction) was used at 3.6 cm pH 4–7 IPG strip using the cup-loading technique kV. Mass data acquisitions were piloted by MassLynx 4.0 (manifold, GE Healthcare Biosciences, Australia). IEF was software (Waters). Nano-LC-MS/MS data were collected conducted using the IPGPhor system (10 min at 150 V, 10 by data-dependent scanning, that is, automated MS to min at 500 V, 10 min at 1,000 V, 1.5 h at 4,000 V, and 4 MS/MS switching. Fragmentation was performed using to 5 h at 8,000 V, total = 50 kVh; GE Healthcare Bio- argon as the collision gas and with a collision energy pro- sciences, Australia). The second dimension was per- file optimised for various mass ranges of ion precursors. formed on 11.5% SDS-PAGE, using the EttanDAlt system Four ion precursors were allowed to be fragmented at the (GE Healthcare Biosciences, Australia). Gels were stained same time. Mass data collected during a NanoLC-MS/MS with Colloidal Brilliant Blue (CBB), and digitised using an analysis were processed automatically with the ProteinL- Image Scanner (Amersham Pharmacia) and the LabScan ynx Process (Waters) module. Data analysis was per- software (v 3.0, Amersham Pharmacia Biotech). Differen- formed with Mascot (Matrix Science Ltd., London, U.K.) tial protein expression analysis was performed using the against the in-house Thiomonas sp. 3As protein database ImageMaster 2D platinum software (v. 6.01, GE Health- with carbamidomethylation (Cys), oxidation (Met), 0.25 care Biosciences, Australia), as previously described [37]. Da mass error and one miss cleavage. All identifications Only spots with a Student's-t value greater than 2 (P value were incorporated into the "InPact" proteomic database less than 0.05) and ratio greater than 2 were analysed. The developed previously http://inpact.u-strasbg.fr/~db/[38]. selected spots were cut from the 2D-gel. Destaining, reduction/alkylation steps by the liquid handling robot Molecular microbiology QuadZ215 (Gilson International, France) and analyses by DNA was extracted and purified from liquid cultures of MALDI-TOF were performed as previously described [37]. pure isolates using the Wizard genomic DNA extraction Tryptic mass searches retained only data with up to one kit (Promega, U.S.A.). The 16S rRNA genes were amplified missed tryptic cleavage and optional methionine oxida- by PCR using the 27f:1492r primer pair [39]. A 743 nt- tion, with mass accuracy limited to 50 ppm. If necessary, long fragment of the rpoA gene of each organism was unidentified proteins were subjected to Nano LC-MS/MS amplified using the rpoAf2a:rpoAr2a primer pair (GGB- analysis. The resulting digest solution was diluted 1:4 into GTGSTCCACGARTAY and GCRAGSACTTCCTTRATYTC, Nano HPLC solvent A (97.9% H2O, 2% ACN and 0.1% respectively). The aoxAf:aoxABr primer pair (TGYACCCA- (v/v) HCOOH). The digested proteins were analysed YATGGGMTGYCC and CSATGGCTTGTTCRGTSASGTA,

Table 3: PCR target and GenBank Accession IDs for strains used in this study.

Strain 16S rpoA aoxAB arsB1 arsB2

3As AM492684a EU339226 EU339209 EU339214 EU339217 Ynys1 AF387302a EU339223 n/d EU339216 n/d WJ68 AY455805a EU339224 EU339213 n/s n/d T. arsenivorans AY950676a EU339231 EU304260a n/d EU339222 T. perometabolis AY455808a EU339230 n/d EU339215 n/d

a Accession IDs from other studies; n/d, no data; n/s, sequence not submitted: the arsB1 and arsB2 sequences obtained with the internal primers were short and therefore were not submitted to the GenBank sequence repository.

Page 10 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

respectively) were used to amplify 1451 nt of the aoxA and aoxB genes, including the short (~27 nt) intragenic region. Additional material The generic arsBf:arsBr primer pair (GGTGTGGAA- CATCGTCTGGAAYGCNAC and CAGGCCGTACACCAC- CAGRTACATNCC, respectively) were designed to amplify Additional file 1 MS (Maldi or MS/MS) identification results of arsenic-induced pro- between 740 and 760 bp of both copies of the arsB gene teins in T. arsenivorans and Thiomonas sp. 3As. Protein profiles in all Thiomonas strains. Following subsequent analysis, expressed in MCSM or m126 media, in the presence and absence of arsB1- and arsB2-specific internal forward and reverse arsenic: detailed results of proteomic and mass spectrometry analyses. primers were designed. The arsB1i2f:arsB1i2r primer pair Click here for file (TGGCGTTCGTGATGGCNTGCGG and CACCG- [http://www.biomedcentral.com/content/supplementary/1471- GAACACCAGCGSRTCYTTRAT, respectively) amplified 2180-9-127-S1.xls] 268 bp of the arsB1 gene, whereas the arsB2i2f:arsB2i1r primer pair (TGGCCGTGGCCTGTTYGCNTTYYT and ACCCAGCCAATACGAAAGGTNGCNGGRTC, respec- tively) amplified 417 bp of the arsB2 gene. Virtual diges- Acknowledgements T. perometabolis was obtained from the Pasteur Institute, Paris, France. The tions of the arsB1 and arsB2 genes of strain 3As suggested authors would like to thank Dr Violaine Bonnefoy and Dr Kevin Hallberg that the two genes should be differentiated by restriction for providing the Thiomonas strains and their invaluable advice on all things fragment length polymorphism (RFLP) analysis using the Thiomonas and Dr Catherine Joulian for her help with functional gene restriction enzyme RsaI. primer design.

Phylogenetic analysis The study was financed by the ANR 07-BLAN-0118 project (Agence Sequences were aligned using the ClustalX alignment pro- Nationale de la Recherche). CGB was supported by a grant from the Uni- gramme [40]. SuperGene analysis was performed by con- versity Louis-Pasteur of Strasbourg. MM was supported by a grant from ANR COBIAS project (PRECODD 2007, Agence Nationale de la Recher- catenating the 16S rRNA and rpoA gene sequences of each che). This work was performed within the framework of the research net- organism, to improve the phylogenetic analysis as pro- work "Arsenic metabolism in Prokaryotes" (GDR2909-CNRS). posed recently [41]. Neighbour-Joining trees were con- structed using ClustalX, with bootstrap values determined References from 1000 replications. Maximum likelihood (ML) trees 1. Abernathy CO, Liu YP, Longfellow D, Aposhian HV, Beck B, Fowler were constructed using the PhyML algorithm [42]. The B, Goyer R, Menzer R, Rossman T, Thompson C, et al.: Arsenic: health effects, mechanisms of actions, and research issues. ModelGenerator programme [43] was used to select the Environ Health Perspect 1999, 107(7):593-597. optimal nucleotide substitution model for ML analysis. 2. Hallberg KB, Johnson DB: Microbiology of a wetland ecosystem constructed to remediate mine drainage from a heavy metal Bootstrap values were determined from 500 replications. mine. Sci Total Environ 2005, 338(1–2):53-66. A list of sequences generated during this study and their 3. Oremland RS, Stolz JF: The ecology of arsenic. Science 2003, GenBank Accession IDs can be found in Table 3. 300(5621):939-944. 4. Casiot C, Morin G, Juillot F, Bruneel O, Personné JC, Leblanc M, Duquesne K, Bonnefoy V, Elbaz-Poulichet F: Bacterial immobiliza- Authors' contributions tion and oxidation of arsenic in acid mine drainage (Car- CGB carried out the physiological and molecular genetic noulès creek, France). Water Res 2003, 37(12):2929-2936. 5. Inskeep WP, Macur RE, Hamamura N, Warelow TP, Ward SA, Santini studies and drafted the manuscript. MM carried out motil- JM: Detection, diversity and expression of aerobic bacterial ity tests, analysed the proteomic data and helped to draft arsenite oxidase genes. Environ Microbiol 2007, 9(4):934-943. 6. Prasad KS, Subramanian V, Paul J: Purification and characteriza- the manuscript. FBB performed the carbon fixation exper- tion of arsenite oxidase from Arthrobacter sp. Biometals 2009 in iments. VK carried out the proteomic experiments. CL-G press. performed the mass spectrometry analyses. DL partici- 7. Ellis PJ, Conrads T, Hille R, Kuhn P: Crystal structure of the 100 kDa arsenite oxidase from Alcaligenes faecalis in two crystal pated in physiological analyses. PB and FA-P conceived of forms at 1.64 A and 2.03 A. Structure 2001, 9(2):125-132. the study, participated in its design and coordination, and 8. Silver S, Phung LT: Genes and enzymes involved in bacterial oxidation and reduction of inorganic arsenic. Appl Environ helped to draft the manuscript. All authors read and com- Microbiol 2005, 71(2):599-608. mented on the manuscript. 9. Muller D, Lièvremont D, Simeonova DD, Hubert JC, Lett MC: Arsenite oxidase aox genes from a metal-resistant beta-pro- teobacterium. J Bacteriol 2003, 185(1):135-141. 10. Santini JM, Hoven RN van den: Molybdenum-containing arsenite oxidase of the chemolithoautotrophic arsenite oxidizer NT- 26. J Bacteriol 2004, 186(6):1614-1619. 11. Lebrun E, Brugna M, Baymann F, Muller D, Lièvremont D, Lett MC, Nitschke W: Arsenite oxidase, an ancient bioenergetic enzyme. Mol Biol Evol 2003, 20(5):686-693. 12. Duquesne K, Lieutaud A, Ratouchniak J, Muller D, Lett MC, Bonnefoy V: Arsenite oxidation by a chemoautotrophic moderately acidophilic Thiomonas sp.: from the strain isolation to the gene study. Environ Microbiol 2008, 10:228-237.

Page 11 of 12 (page number not for citation purposes) BMC Microbiology 2009, 9:127 http://www.biomedcentral.com/1471-2180/9/127

13. Bruneel O, Personné JC, Casiot C, Leblanc M, Elbaz-Poulichet F, terial population: selection, characterization, and perform- Mahler BJ, Le Flèche A, Grimont PA: Mediation of arsenic oxida- ance in reactors. J Appl Microbiol 2002, 93(4):656-667. tion by Thiomonas sp. in acid-mine drainage (Carnoulès, 32. Bryan CG, Hallberg KB, Johnson DB: Mobilisation of metals in France). J Appl Microbiol 2003, 95(3):492-499. mineral tailings at the abandoned São Domingos copper 14. Johnson DB, Hallberg KB: Biogeochemistry of the compost bio- mine (Portugal) by indigenous acidophilic bacteria. Hydromet- reactor components of a composite acid mine drainage pas- allurgy 2006, 83(1–4):184-194. sive remediation system. Sci Total Environ 2005, 338(1–2):81-93. 33. Weeger W, Lièvremont D, Perret M, Lagarde F, Hubert J-C, Leroy M, 15. Coupland K, Battaglia-Brunet F, Hallberg KB, Dictor MC, Garrido F, Lett M-C: Oxidation of arsenite to arsenate by a bacterium Johnson DB: Oxidation of iron, sulfur and arsenic in mine isolated from an aquatic environment. BioMetals 1999, waters and mine wastes: an important role of novel Thi- 12(2):141-149. omonas spp. In Biohydrometallurgy: a sustainable technology in evolution 34. Kolmert Å, Wikström P, Hallberg KB: A fast and simple turbidi- Edited by: Tsezos AHM, Remondaki E. Zografou, Greece: National metric method for the determination of sulfate in sulfate- Technical University of Athens; 2004:639-646. reducing bacterial cultures. J Microbiol Methods 2000, 16. Katayama Y, Uchino Y, Wood AP, Kelly DP: Confirmation of Thi- 41(3):179-184. omonas delicata (formerly Thiobacillus delicatus) as a distinct 35. Miles AA, Misra SS: Estimation of the bactericidal power of the species of the genus Thiomonas Moreira and Amils 1997 with blood. J Hyg (London) 1938, 38:732-749. comments on some species currently assigned to the genus. 36. Bertin P, Benhabiles N, Krin E, Laurent-Winter C, Tendeng C, Turlin Int J Syst Evol Microbiol 2006, 56(Pt 11):2553-2557. E, Thomas A, Danchin A, Brasseur R: The structural and func- 17. Moreira D, Amils R: Phylogeny of Thiobacillus cuprinus and tional organization of H-NS-like proteins is evolutionarily other mixotrophic thiobacilli: proposal for Thiomonas gen. conserved in gram-negative bacteria. Mol Microbiol 1999, nov. Int J Syst Bacteriol 1997, 47(2):522-528. 31(1):319-329. 18. Kelly DP, Uchino Y, Huber H, Amils R, Wood AP: Reassessment of 37. Weiss S, Carapito C, Cleiss J, Koechler S, Turlin E, Coppee J-Y, Hey- the phylogenetic relationships of Thiomonas cuprina. Int J Syst mann M, Kugler V, Stauffert M, Cruveiller S, et al.: Enhanced struc- Evol Microbiol 2007, 57(Pt 11):2720-2724. tural and functional genome elucidation of the arsenite- 19. London J, Rittenberg SC: Thiobacillus perometabolis nov. sp., a oxidizing strain Herminiimonas arsenicoxydans by proteomics non-autotrophic Thiobacillus. Arch Microbiol 1967, data. Biochimie 2009, 91:192-203. 59(1):218-225. 38. Bertin PN, Médigue C, Normand P: Advances in environmental 20. Katayama-Fujimura Y, Kuraishi H: Emendation of Thiobacillus genomics: towards an integrated view of micro-organisms perometabolis London and Rittenberg 1967. Int J Sys Bacteriol and ecosystems. Microbiology 2008, 154(Pt 2):347-359. 1983, 33:650-651. 39. Lane DJ: 16S/23S sequencing. In Nucleic acid techniques in bacterial 21. Battaglia-Brunet F, Joulian C, Garrido F, Dictor MC, Morin D, Coup- systematics Edited by: Goodfellow ESM. New-York: John Wiley and land K, Barrie Johnson D, Hallberg KB, Baranger P: Oxidation of Sons; 1991:115-175. arsenite by Thiomonas strains and characterization of Thi- 40. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG: The omonas arsenivorans sp. nov. Antonie van Leeuwenhoek 2006, CLUSTAL_X windows interface: flexible strategies for mul- 89(1):99-108. tiple sequence alignment aided by quality analysis tools. 22. Hallberg KB, Johnson DB: Novel acidophiles isolated from mod- Nucleic Acids Res 1997, 25(24):4876-4882. erately acidic mine drainage waters. Hydrometallurgy 2003, 41. Gadagkar SR, Rosenberg MS, Kumar S: Inferring species phyloge- 71:139-148. nies from multiple genes: concatenated sequence tree ver- 23. Bodénan F, Baranger P, Piantone P, Lassin A, Azaroual M, Gaucher E, sus consensus gene tree. J Exp Zoolog B Mol Dev Evol 2005, Braibant G: Arsenic behaviour in gold-ore mill tailing, Massif 304(1):64-74. Central, France: hydrogeochemical study and investigation 42. Guindon S, Gascuel O: A simple, fast, and accurate algorithm of in situ redox signatures. Applied Geochemistry 2004, to estimate large phylogenies by maximum likelihood. Syst 19:1785-1800. Biol 2003, 52(5):696-704. 24. Quéméneur M, Heinrich-Salmeron A, Muller D, Lièvremont D, 43. Keane TM, Creevey CJ, Pentony MM, Naughton TJ, McLnerney JO: Jauzein M, Bertin PN, Garrido F, Joulian C: Diversity surveys and Assessment of methods for amino acid matrix selection and evolutionary relationships of aoxB genes in aerobic arsenite- their use on empirical data shows that ad hoc assumptions oxidizing bacteria. Appl Environ Microbiol 2008, 74(14):4567-4573. for choice of matrix are not justified. BMC Evol Biol 2006, 25. Muller D, Médigue C, Koechler S, Barbe V, Barakat M, Talla E, Bon- 6:29-47. nefoy V, Krin E, Arsène-Ploetze F, Carapito C, et al.: A tale of two oxidation states: bacterial colonization of arsenic-rich envi- ronments. PLoS Genet 2007, 3(4):e53. 26. Shock EL, Helgeson HC: Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures: correlation algorithms for ionic species and equation of state predictions to 5 kb and 1000°C. Geo- chimica and Cosmochimica Acta 1988, 52(8):2009-2036. 27. Starkey RL: Precipitation of Ferric Hydrate by Iron Bacteria. Science 1945, 102(2656):532-533. 28. Carapito C, Muller D, Turlin E, Koechler S, Danchin A, Van Dorsse- laer A, Leize-Wagner E, Bertin PN, Lett MC: Identification of genes and proteins involved in the pleiotropic response to Publish with BioMed Central and every arsenic stress in Caenibacter arsenoxydans, a metalloresistant beta-proteobacterium with an unsequenced genome. Bio- scientist can read your work free of charge chimie 2006, 88(6):595-606. "BioMed Central will be the most significant development for 29. Parvatiyar K, Alsabbagh EM, Ochsner UA, Stegemeyer MA, Smulian disseminating the results of biomedical research in our lifetime." AG, Hwang SH, Jackson CR, McDermott TR, Hassett DJ: Global analysis of cellular factors and responses involved in Pseu- Sir Paul Nurse, Cancer Research UK domonas aeruginosa resistance to arsenite. J Bacteriol 2005, Your research papers will be: 187(14):4853-4864. 30. Zhang Y, Ma YF, Qi SW, Meng B, Chaudhry MT, Liu SQ, Liu SJ: available free of charge to the entire biomedical community Responses to arsenate stress by Comamonas sp. strain CNB- peer reviewed and published immediately upon acceptance 1 at genetic and proteomic levels. Microbiology 2007, 153(Pt 11):3713-3721. cited in PubMed and archived on PubMed Central 31. Battaglia-Brunet F, Dictor MC, Garrido F, Crouzet C, Morin D, yours — you keep the copyright Dekeyser K, Clarens M, Baranger P: An arsenic(III)-oxidizing bac- Submit your manuscript here: BioMedcentral http://www.biomedcentral.com/info/publishing_adv.asp

Page 12 of 12 (page number not for citation purposes)